SocialCircle: AI-Powered Relationship Intelligence
For the generation that meets everyone and stays close to no one.
Presentation Deck
01. PROJECT OVERVIEW
Building Relationships Beyond the Match
Designing a friendship-deepening mobile app around a counselor's proprietary 5-tier relationship framework, and validating it with real users. I led this project end-to-end, from discovery and persona development through prototype design, usability testing, and design iterations.
Client
Relationship Counselor
Duration
6 Weeks, Feb 2026 - Mar 2026
Platform
iOS, Android
My Responsibility
Product Strategy, UX Research, UX Design, Usability Testing, Business Model Canvas
Tools Used
Figma Design, Figma Slides, ChatGPT, Google Workspace
02. CLIENT BRIEF
What I Was Asked to Solve?
The client, KB, is a professional relationship counselor with a proprietary 5-tier framework for understanding closeness. She needed a platform that would bring users to her, not replace her.
KB uses a model, Inner Circle, Confidants, Allies, Acquaintances, and Strangers, to help people understand who they are truly investing in and whether that investment is mutual. Her practice is 1-on-1. But most people who need her don't know they need her yet. By the time they seek professional support, friendships have faded, patterns have calcified, and the emotional cost is high.
“
I want something people can use on their own first, so they start doing the work before they ever sit down with me. And when they're ready for more, I want them to come to me.
— KB Relationship Counselor, Client
TWO CONSTRAINTS THAT SHAPED EVERYTHING
Framework First
KB's 5-tier framework had to be the foundation, not an afterthought. Every screen needed to make her methodology feel natural and immediately useful.
No Peer-to-Peer Chat
KB was emphatic: the app must not simulate the support she provides professionally. All real human guidance routes through her. The app's job is reflection, not conversation.
These constraints turned out to be design gifts. They gave me a clear point of view from day one and prevented us from building just another messaging app.
03. DISCOVERY
The Loneliest Generation in History
Before sketching anything, I needed to understand why young adults struggle with friendship after the initial meeting.
WHAT THE MARKET TOLD ME
I mapped the competitive landscape and found something striking: every major social app is optimized for broadcasting to many people or facilitating a first meeting. Not one is designed for what comes after.
Snapchat
~18%
What they do well
Massive reach, Gen Z fluency, Stories format
Where they fall short
Broadcast-first; no friendship depth or nurturing tools
Timeleft
~1%
What they do well
Innovative dinner-matching format
Where they fall short
One-shot experience; no sustained relationship building
Bumble BFF
~8%
What they do well
Friend-matching mechanic, safety features
Where they fall short
Stops at the match, no ongoing relationship layer
Patook
~2%
What they do well
Platonic focus, AI safety filters
Where they fall short
No events, no AI reflection, niche UX
SocialCircle's Gap
First-Mover
The only app focused on deepening existing friendships. No direct competitor. Uncontested category.
White Space: AI coaching + client's 5-tier framework + social circle tracking + local events + pathway to professional counseling
WHAT USERS ARE ACTUALLY DOING
Without a structured tool, people improvised and consistently engaged in the same ineffective behaviors: mentally tracking communication frequency (exhausting, inaccurate), re-reading old messages to interpret intent (anxious), passively mirroring energy (leads to drift), or simply doing nothing until the friendship quietly disappears.
$8.5B
Total Addressable Market
280M
Target Users
(ages 18-35)
11.2%
Market CAGR
2023-2030
No. 1
Uncontested Market Gap
04. DEFINE
Who Did I Design for?
Two personas synthesized from research, the same core problem, and two completely different relationships with it.
Sarah Fisher
21, Female | Computer Science Student
Social Circle
Large outer circle, 2-3 emotionally close connections
Relationship Status
Situationship/exploring dating
"If I'm investing emotionally, I need to know it's mutual."
Bio
Sarah is socially active but emotionally selective. She interacts with many people daily through work and college, but she deeply connected to only a few. She values emotional safety and clarity in relationships but struggles to ask directly where she stands.
Goals
Define relationship status clearly
Reduce emotional ambiguity
Feel secure before investing deeply
Frustrations
No structure to define progression
Messaging apps don't clarify depth
Everything depends on guessing
Needs
Clear communication without awkward confrontation.
Signals of mutual effort.
Emotional reassurance.
Consistency in behavior.
Behaviors
Tracks communication frequency mentally.
Notices drop in effort quickly
matches energy when hurt.
Withdraws silently instead of confronting.
Uses social media to interpret emotional cues.
Percy Jackson
25, Male | Bio Technology Professional
Social Circle
Stable, small-to-medium
Relationship Status
Single/casually dating
"If I feel ignored repeatedly, I just stop trying."
Bio
Percy has a stable social ecosystem, an office circle, and long-term friends. He does not actively seek new connections unless naturally introduced through shared environments. He believes relationships should grow organically, not through forced structure. He values mutual respect and effort.
Goals
Maintain stable friendships
Avoid drama and dependency
Invest where effort is reciprocated
Frustrations
Platforms don't reflect depth
No structured way to track relational health
Effort tracking is subconscious
Needs
Emotional maturity.
Honest but calm communication.
Low-drama environments.
Behaviors
Rarely initiates confrontations.
Relies on pattern recognition to detect change.
Prefers organic deepening over defined stages.
JOURNEY MAP - SARAH FISHER
Scenario: Meeting someone new and navigating emotional progression.
Stage
Initial Contact
Growing Familiarity
Emotional Investment
Ambiguity & Drift
Withdrawal/
Ghosting
Action
Casual chats, light engagement with a new person
Shares details; prioritizes interaction
Opens up emotionally, invests significant time and energy
Re-reads chats, pulls back, mirrors other's energy
Stops initiating; goes quiet or gets ghosted
Thoughts
"They seem interesting. Let's see where this goes."
"We're getting closer. Are we on the same page?"
"This is becoming important. I hope they feel the same."
"Did I do something? Am I reading too much into this?"
"Was it even real? I should have asked sooner."
Emotions
Curious, open
Hopeful, attached
Vulnerable, invested
Anxious, confused
Disappointed, self-doubt
😐
☹️
😊
😄
🙂
Pain Points
Feels pressure to perform match others' social ease
No structured way to know if effort is being matched by the other person
No expectation conversation, assumes mutual investment
No tool to evaluate reciprocity or see patterns; fear of asking directly
Emotional withdrawal carries into next relationship; missed learning opportunity
Opportunities
Confidence-building warm-up prompts help Sarah frame her social goals
Social circle tiers show her where a person sits and how effort compares
AI reflection bot surfaces balance of effort before the drift begins
Structured reciprocity check-in replaces guesswork; KB session if stuck
Post-fade reflection prompt helps Sarahunderstand whathappened and why
The breakdown happens between Stage 3 → Stage 4. The person senses a shift but has no tool to understand or respond to it. That is exactly where SocialCircle intervenes.
05. FRAME
Framing the Design Challenge
“
As a young adult, when I sense a shift in effort or communication in a friendship, I want a clear way to understand the health of that connection — but there is no structured way to evaluate reciprocity.
Problem Statement
How Might We… help young adults invest in new friendships so connections don't fade before they have a chance to deepen?
Two things are notable about this framing. First, it is not about meeting people, every existing solution addresses meeting. I deliberately positioned SocialCircle in the space after that. Second, it is about empowerment, not surveillance. The goal is to help users make informed decisions, not to score or rank their relationships.
06. IDEATE
Exploring Solutions
With KB's constraints defined, I explored what tools would actually move the needle for Sarah and Percy.
OPTIONS I CONSIDERED
OPTIONS 1: REJECTED
Communication Tracker
Log messages, calls, meetups, show frequency over time. Felt cold and transactional. Manual logging would be abandoned quickly.
OPTIONS 2: REJECTED
Journaling Tool
Prompted daily reflections on specific relationships. Closer to KB's methodology but too heavy, users wouldn't open a journal every day for a friendship they're still figuring out.
OPTIONS 3: CHOSEN
Social Circle Visualizer + AI Nudges
A visual map of relationships by closeness tier, with an AI that surfaces reflection questions after social interactions. Lightweight but structured. Maps directly onto KB's framework.
OPTIONS 4: CHOSEN
Events Layer
Surface local events and make it easy to invite a specific circle member. Turns reflection into action. Became the highest-validated feature in testing.
WHY I REJECTED CHAT
At several points I discussed whether in-app messaging would increase engagement. I stress-tested the constraint and concluded: chat would make SocialCircle a messaging app competing with iMessage and WhatsApp, would dilute the reflective purpose, and, most critically, would reduce the need for KB's sessions. The no-chat constraint was not a limitation. It was the product's identity.
07. DESIGN
What I Built and Why
The features, each mapped to a specific user need from research, and each grounded in KB's framework.
Feature 1
Social Circle Tiers: Client's Framework in Action
The home screen organizes connections into KB's five tiers: Inner Circle, Confidants, Allies, Acquaintances, and Strangers. Every time a user adds someone, they must decide where that person actually sits in their life, the first act of reflection.
Why this decision
The visual circle metaphor was important, closeness is spatial, and people think about relationships in terms of proximity. I was uncertain whether users would understand the tier names without explanation and flagged this for testing.
Feature 2
Events & Sharing: Low-Friction Investment
A browsable events tab showing local activities, with the ability to share an event directly with a specific circle member. Research shows shared experiences are one of the most effective ways to deepen a new friendship, SocialCircle removes the coordination friction.
Why this decision
This feature also gave users a reason to open the app regularly, not just when they're in distress about a relationship. I was unsure whether the KB session booking surfaced here would be noticed.
Feature 3
AI Reflection Bot: Structured Reciprocity Check-ins
A chat interface with an AI coach, deliberately labeled "SocialCircle AI Coach", that asks structured reflection questions after social interactions: Who did you spend time with? Did the effort feel balanced? The bot is a coaching tool, not a friend.
Why this decision
This is the digital version of what KB does in session, guided reflection surfacing patterns the user hasn't consciously noticed. My biggest usability risk going into testing: would users know they were talking to a bot?
Feature 4
1-on-1 Session Booking: The Professional Bridge
A booking interface on the home screen and events tab, enabling users to schedule a session with KB directly. When the AI identifies patterns suggesting a user needs more support, it surfaces the booking prompt naturally.
Why this decision
This is the product's ethical spine. SocialCircle brings users to the point where they know they need KB and makes that next step frictionless. I was uncertain whether users would find it without more prominent placement.
08. PROTOTYPE
What I Actually Made
An interactive Figma prototype covering three end-to-end flows across five screen sections.
THE THREE FLOWS TESTED
FLOW
1
Add a Person to My Social Circle
Home screen → tap add button → enter contact → select circle tier → confirm → person appears in circle
FLOW
2
Find an Event and Share It With Someone
Events tab → browse list → tap event → read detail page → tap Share Link → select circle member → confirm
FLOW
3
Chat with the AI Reflection Bot
Chat tab → read opening question → type response → answer follow-ups → mention a circle member → receive reflection summary
Notable design choice: The add button was a floating plus icon on the home screen with no text label. This proved to be a critical discoverability failure in testing, and one of my biggest learning moments.
09. TESTING
What I Put in Front of Users
Five participants. Three tasks. 60 minutes each. No guidance during tasks.
5
Participants tested
18–35
Age range
60 min
Per session
3
Tasks each
THE THREE FLOWS TESTED
FORMAT
Moderated Think-Aloud
In-person and remote via Zoom. Participant drove the device, no pointing, no guiding. Moderator observed silently with timestamped notes.
PROTOCOL
No App Overview First
I gave no overview before tasks began. I wanted to see what users would discover, or fail to discover, entirely on their own. Probing questions only after each task ended.
TASK SCENARIOS
Task 1: "You just met someone named Alex at a local event. Add Alex to whichever circle you think fits best."
Success: Adds a contact and assigns a tier without guidance.
Task 1: "Browse the events in the app, find one that interests you, and share it with someone from my social circle."
Success: Completes the share flow with a circle member.
Task 1: "Browse the events in the app, find one that interests you, and share it with someone from my social circle."
Success: Completes the share flow with a circle member.
10. FINDINGS
What I Observed
My prioritized issues, and several things that confirmed my hypotheses.
WHAT WORKED
DESIRABLE
Circle concept landed instantly
All 5 participants understood placing people in tiers without any explanation. Nobody asked "what is this screen?" They just started deciding where Alex should go.
DESIRABLE
Events was the emotional highlight
Every participant lit up. Multiple stopped mid-task to say "this is actually really useful." The share flow, once inside an event, felt natural and fast.
VIABLE
Nobody missed chat
Not once in 5 sessions did anyone ask "can I message this person?" The constraint KB insisted on was completely invisible to users in practice.
WHAT DIDN'T WORKED
HIGH
Circle tier names caused hesitation
Finding: 3 of 5 participants paused when choosing a tier for Alex. The concept was immediately understood, the vocabulary was the issue. "Confidant" and "Ally" caused confusion; participants weren't sure which represented a closer relationship.
"The circle tiers make sense, I just wasn't sure if Confidant was closer than Ally. I had to guess."
— Participant 4, Age 29
Tier placement is the primary act of reflection in SocialCircle. If users are guessing rather than deciding intentionally, the feature loses its core value.
Fix: Add a one-line description beneath each tier name. Example: "Confidants, people you tell everything to."
HIGH
"Add Person" button was hard to find
Finding: 3 of 5 participants took over 90 seconds to locate the add button. One participant took 2 min 20 sec. The plus icon blended into the circle visualization, discoverability was the barrier, not the flow.
This is a retention risk: a user who can't add someone in the first session may not return for a second.
Fix: Increase button size, add "Add to Circle" text label, reposition to a clearly visible and predictable location.
HIGH
AI bot purpose was not immediately clear
Finding: 4 of 5 participants opened chat and initially assumed they were talking to a real person. Realization took 20–30 seconds. Participants recalibrated quickly, but the first impression was not what I intended.
"I thought I was messaging someone. Then I realized, oh, it's like an AI asking me questions about my friendships. That's actually cool."
— Participant 3, Age 27
Fix: Add a persistent label: "SocialCircle AI Coach — reflecting on ymy social life." One-sentence explainer on first open. Always visible.
KB session booking was invisible
Finding: No participant found KB booking independently during any task. When probed in debrief, they responded enthusiastically, but had walked right past it. This is a placement problem, not a concept problem.
Fix: Increase visual prominence on home screen. Add contextual bot prompt when user expresses significant uncertainty during a reflection session.
MEDIUM
WHAT PARTICIPANTS SAID
"I love the idea of having a map of who I'm actually investing time in. It's like a CRM for friendships."
— Participant 2, Age 24
"Finding an event and being able to invite someone right there, that's the feature that would make me download it."
— Participant 1, Age 22
"I haven't seen anything quite like this."
— 4 of 5 Participants
"If I could actually book a session with someone who knows about relationships, I'd pay for that."
— Multiple participants
11. VALIDATION
Validated as Desirable, Feasible & Viable
Desirable
Users genuinely want this
5/5 found concept highly desirable
Events = #1 reason to keep using the app
Circle tracking felt novel and useful
Strong resonance with recent movers
Feasible
The prototype works end-to-end
All 3 Figma flows validated
Events tab found in <5 sec by all users
Share flow described as fast and natural
3 prioritized fixes scoped and ready
Viable
Clear willingness to pay
3 of 5 would pay for KB coaching
Freemium cap accepted as fair trigger
No direct competitor confirmed
Strong B2B signal from movers segment
12. ITERATIONS
Three Prioritized Design Changes
My prioritized issues, and several things that confirmed my hypotheses.
1
Add tier descriptions to add-person screen
Before: Each tier displayed its name only, "Confidants," "Allies", with no supporting context.
After: Participants accepted the 5-person limit as a clear and fair payment trigger
2
Redesign the "Add to Circle" entry point
Before: Floating plus icon, low contrast, no label. 3 of 5 users took over 90 seconds to find it.
After: Labeled button ("Add to Circle"), larger tap target, repositioned to a prominently visible location.
3
Label the AI bot clearly on chat open
Before: Chat opens with the bot's first question, no context. 4 of 5 participants expected a real person.
After: Persistent header, "SocialCircle AI Coach, reflecting on my social life." One-sentence explainer on first open only.
13. REFLECTION
What I Learned
Post-project reflection on process, decisions, and surprises.
WHAT WORKED
Starting with constraints paid off. The two constraints felt limiting at first but gave us a clearer product identity than I'd have arrived at without them.
Testing early with real scenarios revealed the right problems. I didn't test whether users "liked" the app, I gave each participant a specific person to add and watched what happened.
The journey map paid dividends. Stage 3 → Stage 4 became a north star for every feature decision throughout the project.
WHAT I'D DO DIFFERENTLY
Test tier naming earlier. A quick card sort in discovery would have caught the Confidants/Allies ambiguity before I built the prototype.
Bring KB into the design process more actively. I interpreted her framework through my own lens, more collaborative sessions, especially around AI bot questions, would have strengthened the product.
Design a specific task around KB booking. I never gave participants a task centered on discovering the booking feature, so I got weak data on the most important revenue interaction.
WHAT SURPRISED ME
Events was far more exciting than I expected. I designed it as a supporting feature. In testing, it became the primary reason participants said they'd download the app.
Users found the metaphor before I gave it to them. "It's like a CRM for friendships", Participant 2 said this within minutes, without prompting. When a user finds the metaphor for your product on their own, you've built something that makes sense.
Nobody missed chat. Five sessions. Zero mentions. The constraint that felt hardest to defend was completely invisible to users in practice.
14. NEXT STEPS
Where the Design Goes Next
Iteration 2 Prototype
Implement all 4 prioritized fixes, test two layout variations for the add button
Design the tier description UI in context, test two copy approaches
Design the persistent bot header and first-open explainer
Increase KB booking prominence, test two placement variations
Second Round of Testing
Retest Tasks 1 and 3, verify fixes resolved the identified issues
Add Task 4: specifically targeting KB session booking discovery
Expand participant pool to include users who recently relocated, my strongest B2B signal
For Client Specifically
Co-design AI reflection questions with KB, she should author the tone and substance of what the bot asks
Define the escalation protocol: what patterns suggest a user needs human support?
Establish how session insights (anonymized) could inform AI coaching logic over time
A strategic choice to protect the client's professional positioning, maintain the app's reflective purpose, and ensure that human intimacy remains the exclusive domain of the client's counseling practice.
